[3P6-OS-9b] OS-9

[3P6-OS-9b-03] Simulating situation-appropriate emotion in collaborative learning using cognitive models

〇Ruiki Kawaji¹, Junya Morita¹, Hirotaka Osawa² (1. Shizuoka University, 2. Keio University)

Keywords:Cognitive model, Hanabi, ACT-R, Emotion, Collaborate

Emotions are considered essential for human communication, and in recent years, research on Affective Computing, which aims to give emotions to computers that interact with humans, has been developing. However, most research on emotions focuses on detecting emotions from body data or generating emotions through machine learning, and there is little research focusing on the cognitive role of emotions. Therefore, we aimed to observe the role that emotions play during human communication through simulations using a cognitive model. In this study, we used the cognitive architecture ACT-R to build a model in which behavior changes through case-based learning depending on emotions. Simulations of a cooperative task using the cooperative game ”Hanabi” showed that appropriate emotional states exist depending on the situation, and that emotions can be adjusted to an appropriate state by fluctuating.